Backup Policy

This Backup Policy describes the operational practices of Rollin Servicos Digitais e Tecnologia LTDA (CNPJ 64.204.851/0001-39), hereinafter "Rollin Host", regarding the creation, retention, storage and restoration of backup copies for hosting services, virtual servers (VPS), dedicated servers and other infrastructure solutions contracted by customers ("Customer").

This policy is incorporated into the Terms of Service and must be read alongside the Transparent SLA and the Privacy Policy. In the event of any conflict, the instrument most specific to the contracted service prevails.

• Backup: a copy of the Customer's service data at a given point in time (snapshot or agentless), created by Rollin Host according to the contracted plan.
• VPS/dedicated snapshot: an image of the virtual disk at a specific instant, captured at the hypervisor level.
• Shared hosting/reseller backup: a copy of the site directory files and associated databases.
• Backup window: the time of day (typically overnight, Brasilia time) when the backup routine runs.
• RPO (Recovery Point Objective): the maximum amount of data that can be lost, measured in time (interval between backups). The effective RPO varies by plan.
• RTO (Recovery Time Objective): the estimated time to complete a restoration after an approved request.
• Retention: the period during which a backup remains available for restoration.

3.1 Shared hosting and reseller

We perform automatic daily backups of:

• Files in the public_html directory and other DirectAdmin user directories.
• MySQL/MariaDB databases associated with the account.
• Email accounts and settings (POP/IMAP mailboxes, filters, redirects).
• Domain settings (managed DNS zones, redirects, subdomains).

3.2 VPS, cloud servers, GPU and dedicated servers

These do not include automatic backup. Backup is the sole responsibility of the Customer, who has root access and full control of the server. We recommend independent routines (database dumps, synchronization to external storage such as AWS S3 or Backblaze B2) and periodic restoration testing.

3.3 VPS snapshots (on demand)

The Customer may create on-demand snapshots of their VPS from the control panel — a point-in-time image of the disk at the moment of capture. Snapshots are a convenience feature (a restore point before a change) and do not replace backups: they reside on the same infrastructure as the server, are limited in quantity per plan, and expire automatically after 30 days. Creating, managing and using snapshots is done by the Customer.

3.4 Automatic Backup add-on (VPS)

For VPS, the Customer may purchase the Automatic Backup add-on, which performs periodic automatic copies of the server with self-service recovery. Specific pricing, frequency and retention are presented at purchase. Without the add-on, section 3.2 applies in full.

4.1 Standard window

The standard execution window is 00:00 to 06:00 (Brasilia time). During this period, there may be slight server I/O degradation. Heavy database workloads or ETL jobs scheduled by the Customer should avoid this window.

4.2 Effective RPO

"Rolling daily" means that once the maximum number of points is reached, the oldest backup is discarded to make room for the most recent one. No standard plan includes indefinite retention.

Hosting backups are stored on infrastructure that is external and geographically isolated from the production server, with encryption at rest and in transit. VPS snapshots, by their nature, reside on the same hypervisor infrastructure — one more reason they do not replace backups.

Backup data remains under a jurisdiction compatible with the Terms of Service and the Privacy Policy. The specific storage location may be disclosed upon request to the DPO (dpo@rollinhost.com.br).

7.1 Who may request

Restoration may only be requested by the registered account holder or a formally designated technical contact, via panel authentication or a ticket submitted from the registered email address. Requests through informal channels (WhatsApp, social media) do not replace the official ticket.

7.2 Procedure

1. The Customer opens a ticket at painel.rollinhost.com.br specifying: the target date/time of the restore point, scope (full account, specific file, specific database) and destination (overwrite production or staging environment).
2. The technical team validates the availability of the requested point and the impact of the operation.
3. The Customer confirms the operation in writing, acknowledging that overwrites are irreversible (the version prior to restoration is not preserved unless a pre-restoration backup is requested).
4. Restoration is performed in the first available technical window.
5. Ticket confirmation with checksum where applicable.

7.3 Reference RTO

RTOs are best-effort targets and depend on data volume. They do not constitute an uptime guarantee; for that, see the Transparent SLA.

7.4 Free assisted restoration

Every active account is entitled to 1 (one) free assisted restoration per month, counted per completed request. Additional restorations may be charged per the current rate card, except when the failure is attributable to a Rollin Host incident.

The following situations are not covered by the standard automatic backup:

• Accounts suspended for violation of the Terms of Service or Acceptable Use Policy — during the suspension period, new backups may cease to be generated.
• Accounts exceeding 50 GB on shared hosting/reseller plans. Above that volume, Rollin Host may halt automatic inclusion and recommend a dedicated backup solution.
• VPS, cloud, GPU and dedicated servers without the Automatic Backup add-on — those services carry no automatic backup by Rollin Host (section 3.2).
• Databases larger than 2 GB on shared hosting — without transactional consistency guarantees.
• Temporary data: caches, PHP sessions, queues, files in /tmp, rotating logs, pending upload files.
• Data on external CDNs, third-party object buckets (S3, R2, Spaces) and other providers outside the Rollin Host infrastructure.
• Internal backups created by the Customer and stored as files on the server (we do not "backup the backup").
• Commercial software, licenses and product keys (must be preserved by the Customer).
• Root-level configurations on VPS and dedicated servers (Customer's responsibility, section 3.2).

Backup is a shared responsibility. Rollin Host provides the operational safety net described in this policy; the Customer must maintain their own independent routine, especially for:

• Mission-critical data (systems that cannot tolerate more than 24 hours of data loss).
• Legal obligations requiring minimum retention periods beyond those provided in this policy (accounting, tax, regulatory).
• Applications with contractual RPO/RTO commitments to the Customer's own end users.

We recommend the Customer: (a) run scheduled database dumps; (b) sync code versioning to an external repository (GitHub, GitLab, Bitbucket); (c) export backups to off-site storage (object storage, own NAS, another provider); (d) test restoration periodically.

Rollin Host is not liable for data loss resulting from:

• Fortuitous events or force majeure, as defined in art. 393 of the Brazilian Civil Code.
• Acts or omissions of the Customer (accidental deletion outside the retention window, credential compromise, misconfiguration).
• Failure of third parties not controlled by Rollin Host (domain registrar, external CDN, integrations).
• Requests for permanent deletion confirmed in writing by the Customer.

In the event of service cancellation, account closure or non-payment resulting in termination, the associated data and backups follow this schedule:

Once the above periods have elapsed, deletion is permanent and irreversible. The Customer is solely responsible for exporting their data before account closure.

Where backup data contains personal data belonging to third parties, Rollin Host acts as a data processor under the terms of art. 5, VII, of Law No. 13,709/2018 (LGPD), and complies with the obligations set out in the Data Processing Addendum (DPA).

Erasure requests from data subjects (right to be forgotten) are carried out in both production and available backups, subject to technical rotation schedules. Where immediate erasure from a backup is technically unfeasible (e.g., append-only media), we apply the principle of logical quarantine: the record is flagged as inactive, isolated, and physically deleted in the next rotation cycle.

We may amend this policy at any time. Material changes (those that reduce the Customer's rights, shorten retention periods or alter RPO/RTO commitments) will be communicated with 30 days' notice by email and within the control panel. Non-material changes (textual corrections, reorganization) take effect upon publication.

Questions about this policy or requests for restoration:

• Technical support (tickets): painel.rollinhost.com.br
• Contractual inquiries: juridico@rollinhost.com.br
• LGPD rights (DPO): dpo@rollinhost.com.br